ChanServ changed the topic of #asahi-alt to: Asahi Linux: porting Linux to Apple Silicon macs | User-contributed/unofficial distribution ports | Logs: https://alx.sh/l/asahi-alt
swaggie has joined #asahi-alt
ChanServ changed the topic of #asahi-alt to: Asahi Linux: porting Linux to Apple Silicon macs | User-contributed/unofficial distribution ports | Logs: https://alx.sh/l/asahi-alt
<ashi>
Interestingly running MESA_GLES_VERSION_OVERRIDE=3.0 firefox to force WebGL2 (I know it is not ready) thirdroom.io which is much more complex does run now ;)
Dcow has quit [Ping timeout: 480 seconds]
ashi has quit [Read error: No route to host]
ashi has joined #asahi-alt
ashi has quit [Quit: Leaving]
Dcow has joined #asahi-alt
Dcow has quit [Remote host closed the connection]
Dcow has joined #asahi-alt
Dcow has quit [Remote host closed the connection]
Dcow has joined #asahi-alt
Dcow has quit [Remote host closed the connection]
Dcow has joined #asahi-alt
Dcow has quit [Ping timeout: 480 seconds]
Dcow has joined #asahi-alt
swaggie has quit [Remote host closed the connection]
swaggie has joined #asahi-alt
Dcow has quit [Ping timeout: 480 seconds]
jamespmo_ has joined #asahi-alt
jamespmorgan has quit [Ping timeout: 480 seconds]
swaggie has quit [Ping timeout: 480 seconds]
jacksonchen666_ has joined #asahi-alt
jacksonchen666 has quit [Ping timeout: 480 seconds]
<ncopa>
which means that the kernel modules needs to be included in the initramfs image
<ncopa>
does the disk driver need firmware?
Dcow has quit [Ping timeout: 480 seconds]
<sven>
no
<sven>
or, well, the firmware is pre-loaded by iboot and we can’t touch it anyway
<jannau>
ncopa: how are you booting the kernel? the ideal setup is that the boot loader appends firmware.cpio to the initramfs
<jannau>
then the firmware is available in /vendorfw and you don't need the nvme for loading modules which need firmware
<ncopa>
currently via grub, but the idea was to skip the grub step and boot kernel as efi directly
<ncopa>
so, if I understand correctly, the idea is to copy the firwmare to the rootfs every boot?
<j`ey>
a tmpfs
<ncopa>
and the reasoning is so it get firmware updates?
<jannau>
that's one reason, the other reasons are supporting immutable root fs like fedora silverblue and preventing unintended distribution of the non-distributable apple firmware
Dcow has joined #asahi-alt
<ncopa>
so i shoudn't copy the firmware to the linux partition's /lib/firmware/vendor
swaggie has joined #asahi-alt
<jannau>
I think the tmpfs solution is preferred but copying is ok as long as you don't intend to image the rootfs and distribute those images
swaggie has quit [Ping timeout: 480 seconds]
Dcow has quit [Ping timeout: 480 seconds]
<mps>
I think for end users copying firmware to /lib/firmware is quite fine
<mps>
at the end is is already on ESP
<mps>
it is*
<mps>
for example I copy it to 'my' rescue usb and I care to not 'distribute' it
<mps>
with apple we have to be careful anyway
swaggie has joined #asahi-alt
swaggie has quit [Ping timeout: 480 seconds]
Dcow has joined #asahi-alt
Dcow has quit [Ping timeout: 480 seconds]
Dcow has joined #asahi-alt
swaggie has joined #asahi-alt
Dcow has quit [Ping timeout: 480 seconds]
swaggie has quit [Ping timeout: 480 seconds]
swaggie has joined #asahi-alt
<ncopa>
xorg+xfce didnt work very well. seems like keyboard or touchpad is not working
Dcow has joined #asahi-alt
swaggie has quit [Ping timeout: 480 seconds]
<ncopa>
does libinput need any patches?
<ncopa>
or xf86-input-libinput
Dcow has quit [Ping timeout: 480 seconds]
<mps>
ncopa: I tested xfce a week ago, it works fine
<ncopa>
so im missing something
<mps>
but maybe you should create /etc/X11/xorg.conf.d/ libinput.conf file
<mps>
ncopa: alpine works from the beginning without any patches in userspace
<mps>
ncopa: a shameless self promotion (again), alpine was first distro used as daily driver on apple silicon, you have to be proud
<mps>
yes, sometimes I patched some userspace programs but just for playing with some features or options
<mps>
ncopa: did you installed eudev or libudev-zero, good to have with xorg
<ncopa>
eudev
<mps>
then it should work out-of-the-box. I insalled eudev with xfce intentionaly to test it for axboe
<ncopa>
it doesnt work for me. didnt help with the 50-libinput.conf
<ncopa>
i guess i need ssh connection to debug this
<mps>
could you something about libinput in /var/log/Xorg.0.log
<mps>
wait, does keyboard works on console
<mps>
see something*
<ncopa>
keyboard works in console
<ncopa>
i need to reboot and connect via ssh
<mps>
strange, of all xf86 pkgs only xf86-input-libinput is needed
<mps>
xf86 drivers I mean
Ry_Darcy has joined #asahi-alt
<ncopa>
ok, i think i figured out what was wrong
<ncopa>
udev-trigger didnt run
<mps>
ah, good you find it
<mps>
(libudev-zero is the rescue ;-) )
<Ry_Darcy>
§Glanzmann just tried a new install of your Debian. After the install (successful), power recycle and reboot, I am being dropped to a GRUB prompt. Anything changed recently?
<ncopa>
typing from the macbook now
Ry_Darcy has left #asahi-alt [#asahi-alt]
<mps>
\o/
<mps>
ncopa: here is my video from Jan 24 2022 of booting alpine on m1 mbp (j293 model)
<ncopa>
it gives the error "/boot/efi/m1n1 does not exist, is /boot/efi mounted?"
<ncopa>
because I mounted the ESP as /boot
<j`ey>
it's needed to keep the device trees in sync
<mps>
ncopa: yes, it is needed, and you should have ESP mounted under /boot/efi
<ncopa>
i cannot have my ESP mounted as /boot? and skip grub?
<ncopa>
it booted just fine for me so far
<mps>
as j`ey told it updated dtbs, but also m1n1 version defined in 'depends' field in APKBUILD
<mps>
it could be added to /etc/fstab
<mps>
'/dev/nvme0n1p4 /boot/efi auto auto' as example
<mps>
ofc, LABEL or PARTUUID could be used instead of /dev/nvme0n1pX
<mps>
but LABEL will not work with busybox mount applet
<ncopa>
i dont think there are any techincal limitations that prevents us to mount /boot on ESP
<mps>
it is VFAT, that is stopper for me
<mps>
so, I mount it under /boot/efi on the rootfs
<ncopa>
and we should not update it via post-install. we should update it via trigger, so that it gets updated if either m1n1, kernel or u-boot-asahi is updated
<ncopa>
not only on kernel updates
iaguis has quit [Quit: leaving]
<ncopa>
and i think m1n1 package should own the trigger script
<mps>
post-install is there to update proper dtbs because I have few kernel flavors and every one have its dtbs
<mps>
I'm not sure will we have more kernel flavors for alpine
<mps>
if not then it could be simplified ofc
<mps>
ncopa: I'm not against change to make it trigger for m1n1, u-boot and kernel
<ncopa>
what uses m1n1/boot.bin?
<mps>
but not sure is it good idea to have kernel, initramfs, dtbs and grub on vfat
<mps>
ncopa: this is second stage bootloader
<mps>
first stage m1n1 on stub partition loads this second stage
<mps>
m1n1/boot.bin is the m1n1+dtbs+u-boot
Dcow has joined #asahi-alt
Dcow has quit [Ping timeout: 480 seconds]
Glanzmann has joined #asahi-alt
Dcow has joined #asahi-alt
<Glanzmann>
Ry_Darcy: Thank you for letting me know. New installs since 12th December are broken. You're the first one to complain. Anyway, I fixed it. The issue that I was deleting a function that I thought was obsolete, but wasn't. Anyway fixed, pushed, uploaded and tested.
Glanzmann has quit [Quit: back to sleep]
Dcow has quit [Ping timeout: 480 seconds]
<steev>
dang, i keep missing glanzmann :(
<j`ey>
steev: he checks the logs, so just leave a message
<steev>
oh good
<steev>
gonna check what his fixes were today first to see if that fixes my issue as well
<steev>
ah, it does
<steev>
will keep that in mind when i noticed something broken like that
Dcow has joined #asahi-alt
Dcow has quit [Ping timeout: 480 seconds]
<mps>
some of you maybe remember that I destroyed display on my old m1 mpb about half year ago. guess what, I took it from garage and tried to power on, on my surprise display now works
<mps>
\o/
<mps>
not sure how long it will work but I hope for the best
<mps>
s/mpb/mbp/
<steev>
oh nice
<mps>
so I hope i will have spare machine to experiment with and not always ask daughter to take her